database-rds-devops
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- Command Execution: The skill utilizes standard AWS CLI commands to retrieve database cluster configurations and CloudWatch metrics. These operations are essential for health monitoring and are performed within the agent's authorized scope.
- External Downloads: The setup instructions for the optional MCP server reference a script from
astral.shto install theuvpackage manager. This is a standard installation method for well-known developer tooling used to manage the skill's dependencies. - Database Security: The skill is strictly read-only and interacts with database internals through predefined SQL queries. Credentials for database access are retrieved securely from AWS Secrets Manager, ensuring that sensitive information is not exposed in the agent's context or output.
- Indirect Prompt Injection Surface: The skill analyzes database logs (Ingestion point: CloudWatch Logs). While it lacks explicit boundary markers for this untrusted data, its capabilities are restricted to read-only diagnostics (Capability inventory: AWS CLI, MCP SELECT queries) and it uses predefined log patterns for analysis (Sanitization: pattern matching filters).
Audit Metadata